Sample-minded: scholars of color support new counting strategy for Census 2000

Dr. Robert B. Hill knows the consequences that result when the
national census undercounts a community’s population. A locality stands
to lose millions of dollars in federal funding for social services, job
training education, and other programs if the Census Bureau undercounts
the locality’s population, he explains.

“We were making progress on census accuracy, but in 1990, we had a
major undercount.” Hill said. “As a result, a number of mayors sued the
Census Bureau for miscounting their city populations.”

Hill, who is director of the Institute for Urban Research at Morgan
State University in Baltimore, Maryland, advises the Census Bureau on
gathering statistical information about African Americans. He is
particularly concerned about the accuracy of the upcoming census
because the Census Bureau undercounted African Americans and Hispanics
at five times the rate that it missed Whites in the 1990 count.

As a member of the Census Bureau’s African American advisory
committee, Hill supports the bureau’s plans to use scientific sampling
to improve the accuracy of the census in the year 2000. Sampling is a
statistical technique that draws conclusions about a population from a
pool within the target population. The sociologist is among a growing
number of statisticians, researchers, civil rights activists, public
officials, and minority academics who support the use of scientific
sampling in the upcoming census.

The proposal to use sampling, however, has run into opposition from
congressional Republicans, who have vowed to block its use in Census
2000.

Census Bureau officials say sampling would be introduced into the
census after traditional enumeration efforts have been made to reach
all households. Typically, questionnaires are mailed to all American
households. As in a traditional census, enumerators would be sent out
to count the households who failed to send questionnaires back to the
Bureau. Sampling would occur to estimate the population of the people
who are the hardest to reach.

A Presidential Endorsement

Earlier this month, the proposed use of scientific sampling in
Census 2000 made front page headlines on newspapers nationwide when
President Bill Clinton endorsed the idea at a community meeting in
Houston, Texas, with minority leaders and academics.

“It is estimated that if we use good statistical sampling
supplemented by what are called quality checks, where you go out into
selected neighborhoods and actually count heads to make sure that the
sampling is working, that we can cut the error rate to a tenth of a
percent….in the next sample, we would miss — out of a country of
nearly 300 million people by then — only 300,000, as opposed to eight
million in the 1990 Census,” Clinton told the Houston group.

Dr. Tatcho Mindiola, the director for the Center for Mexican
American Studies and an associate professor of sociology at the
University of Houston, informed the president that the community center
meeting was occurring “in an area that has all of the characteristics
of a geographical component that has traditionally been undercounted in
the census.

“You’re at a center that’s located in the low-income area. You’re
at a center that’s located in a predominately Hispanic area of Houston,
an area where there’s a preponderance of immigrants. You’re in an area
where there’s an unusual number of children. You’re in an area where
Spanish is spoken almost as much as English,” Mindiola told Clinton at
the meeting.

Mindiola then praised the president for endorsing census sampling.

“I was pleased to hear you endorse the statistical sampling
technique, which will lead to, in my estimation, a better count of
areas like this particular area where we’re at this morning,” Mindiola
said.

Dr. Sanders Anderson Jr., an associate professor of political
science at Texas Southern University in Houston, said census data is
critical to the research he conducts as a political scientist.

“When you look at the numbers that are included in the census, it
tells us all kinds of information about the American people, It tells
us the progress people have made with regard to their income status. It
tells us where people are by location,” Anderson said.

Dr. Michael T. Nettles, director of the Frederick D. Patterson
Research Institute at the The College Fund/UNCF, said census sampling
is a good idea but planning the sampling will require considerable
effort and sophistication.

“The concept is good; the devil here is in the details. The concept
alone doesn’t ensure that sampling designs will be adequate,” Nettles
warned.

The accuracy of the cenSUS will also directly affect federal higher
education funding since funding formulas are based on population data.
Hill has observed that student leaders nationwide are already beginning
to mobilize to convey this message to their peers who might not
otherwise make the connection between the census and higher education
funding.

Plans to use scientific sampling in Census 2000 developed after
Congress, in the aftermath of the 1990 census, requested that the
bureau accomplish two goals. First, to reduce the undercount — which
increased from 1.2 percent of the population in 1980 to 1.8 percent of
the population in 1990, Congress was specifically interested in
eliminating the high undercount of groups such as African Americans and
Hispanics. Second, Congress charged the Census Bureau with constraining
the cost of the 2000 Census.

A Legislative Battle

Opponents of census sampling have filed two lawsuits to block its
use in 2000. U.S. Representative Bob Barr (R-Ga.) is a plaintiff in a
lawsuit filed last February in Virginia to prohibit census sampling.

“Perhaps someone should explain to the president that the job of
the Census Bureau is to count people, not engage in bogus sampling, ”
he said. “The Constitution uses the words `actual enumeration,’ and not
`statistical sampling,’ `puffed numbers,’ `best guess,’ or some other
such nonsense, for a very clear reason.”

Robert Mallet, deputy secretary of the U.S. Department of Commerce,
said even if the courts allow census sampling, the proponents of
sampling should expect a legislative battle in Congress.

A final decision is required by early next year, according to Commerce department officials.

“The professionals at the Census Bureau at the Department of
Commerce have indicated they absolutely need a decision by February of
next year — absolutely — because we have to do a tremendous amount of
planning, and that’s our drop dead date,” Mallett told news reporters
in Houston.

Dr. David Bositis, a senior research associate at the joint Center
of Political and Economic Studies in Washington, said Republican
opposition to census sampling is based partly on the unstated belief
that a more minority–friendly census will lead to the creation of
additional majority-minority congressional districts, and thus lead to
more Democrats in Congress. However, lie believes the electoral fears
of Republicans are exaggerated and that the GOP would benefit in the
long run with Black and Latino voters by not taking such a hostile
stance on census sampling.

The best way to convince Republican opponents to drop their
opposition to census sampling is to “persuade them that [census
sampling] will not hurt them,” Bositis said.

Hill and Nettles said it’s important that proponents of sampling
make the effort to educate the public about census sampling. Hill added
that public awareness and input can make a difference in the outcome if
Congress has final say on the dispute.

“I think that more pressure can be brought upon Congress. A public
debate can have impact on the direction of this issue,” Hill said.

COPYRIGHT 1998 Cox, Matthews & Associates



© Copyright 2005 by DiverseEducation.com